Probability density distillation with generative adversarial networks for high-quality parallel waveform generation

Abstract

This paper proposes an effective probability density distillation (PDD) algorithm for WaveNet-based parallel waveform generation (PWG) systems. Recently proposed teacher-student frameworks in the PWG system have successfully achieved a real-time generation of speech signals. However, the difficulties optimizing the PDD criteria without auxiliary losses result in quality degradation of synthesized speech. To generate more natural speech signals within the teacher-student framework, we propose a novel optimization criterion based on generative adversarial networks (GANs). In the proposed method, the inverse autoregressive flow-based student model is incorporated as a generator in the GAN framework, and jointly optimized by the PDD mechanism with the proposed adversarial learning method. As this process encourages the student to model the distribution of realistic speech waveform, the perceptual quality of the synthesized speech becomes much more natural. Our experimental results verify that the PWG systems with the proposed method outperform both those using conventional approaches, and also autoregressive generation systems with a well-trained teacher WaveNet.

arXiv commentAccepted to the conference of INTERSPEECH 2019

References in corpus (5)

• WaveNet: A Generative Model for Raw Audio
• Fast Spectrogram Inversion using Multi-head Convolutional Neural Networks
• NSML: Meet the MLaaS platform with a real-world case study
• Generative adversarial network-based glottal waveform model for statistical parametric speech synthesis
• Effective parameter estimation methods for an ExcitNet model in generative text-to-speech systems

Cited by in corpus (7)

• WaveFlow: A Compact Flow-based Model for Raw Audio
• On Leveraging Pretrained GANs for Generation with Limited Data
• End-to-End Adversarial Text-to-Speech
• Multi-band MelGAN: Faster Waveform Generation for High-Quality Text-to-Speech
• FeatherWave: An efficient high-fidelity neural vocoder with multi-band linear prediction
• GAN Memory with No Forgetting
• Parallel waveform synthesis based on generative adversarial networks with voicing-aware conditional discriminators